Post on 19-Oct-2020
transcript
537
APPENDIX A. PERFORMANCE OF REINFORCED SOIL STRUCTURES DURING EARTHQUAKES
Past earthquakes have provided numerous case studies of reinforced soil wall performance
under dynamic loading. These cases have expanded our knowledge and increased the
confidence in this type of retaining systems. In general reinforced soil structures have
performed well in earthquakes. Numerous cases have been reported where reinforced soil
structure performance in major earthquakes have been documented (Chen et al. 2000; Collin
et al. 1992; Eliahu and Watt 1991; Frankenberger et al. 1996; Fukuda and Tajiri 1994; Huang
2000; Kobayashi et al. 1996; Kramer et al. 2001; Kutter et al. 1990; Ling et al. 2001;
Nishimura et al. 1996; Sandri 1994, 1997; Sitar et al. 1997; Stewart et al. 1994a; Stewart et
al. 1994b; Tatsuoka et al. 1995, 1996a; Tatsuoka et al. 1996b; Tatsuoka et al. 1998; White
and Holtz 1996).
LOMA PRIETA EARTHQUAKE, CALIFORNIA, 1989
Numerous reinforced soil walls and slopes were located within the affected area of the 1989
Loma Prieta Earthquake (Collin et al. 1992; Eliahu and Watt 1991; Kutter et al. 1990). The
earthquake had a magnitude 7.1 and peak ground accelerations up to 0.6g were measured.
Several geosynthetic reinforced structures were identified and investigated. Of those
structures reported the highest wall and slope were 5.5 meters and 24.5 meters respectively.
These structures were built with geogrid reinforcements. Reinforcements were wrapped-
around as wall facing in most of the structures. In some of the structures segmental masonry
blocks were used as facing elements. No direct measurements were available at any of these
sites except one. No signs of damage or cracking were observed. One of the reinforced slopes
(La Honda Slope) had an inclinometer and it was possible to estimate the earthquake induced
lateral displacements. Cross section of this slope is shown in Figure A-1. Inclinometer
measurements were made before the earthquake to monitor the slope as shown in Figure A-2.
Comparison of pre- and post-earthquake inclinometer measurements indicate that top of the
slope deformed about 2 cm laterally, corresponding to 0.2% of the slope height.
538
Loma Prieta Earthquake was particularly important to demonstrate the seismic performance
of geogrid reinforced soil structures, because it was the first major seismic event since the
use of structural geogrids initiated in the early 1980s.
Numerous walls built with Reinforced Earth technology (i.e. steel strips and concrete facing
panels) were also investigated for their performance in the Loma Prieta Earthquake
(Reinforced Earth Co. 1998). Twenty Reinforced Earth walls at 9 sites, the highest one being
20 meters, were inspected. No evidence of damage was identified in any of these structures.
Thirty-four reinforced soil structures were reported (Kutter et al. 1990). These included a
variety of construction technologies from steel strips reinforcements, tire-anchor timber walls
to soil nailing. Significant damage was observed at 4.5 meter high tire-anchor timber wall.
However, no details are available from any of these cases. Therefore, it is not possible to
identify if any of these cases are of the ones reported in the other studies described above.
Eight soil nailed excavations were identified within the affected area. Tallest of these
structures was 9.8 meters and all of these structures performed very well, showing no signs
of distress (Felio et al. 1990; Vucetic et al. 1998).
NORTHRIDGE EARTHQUAKE, CALIFORNIA, 1994
Northridge Earthquake was a magnitude 6.7 event with peak ground accelerations reaching
0.9g at locations close to the source of energy release. Northridge Earthquake was an
important event due to its relatively high vertical acceleration levels (Stewart et al. 1994a).
Numerous cases of reinforced soil structure performance were documented (Bathurst and Cai
1995; Sandri 1994, 1997; Stewart et al. 1994a; Stewart et al. 1994b; White and Holtz 1996).
In general reinforced soil structures exhibited excellent performance.
Of those structures reported the highest wall and slope were 11.6 meters and 24.5 meters
respectively. These structures were built with geogrid reinforcements and segmental masonry
blocks facing elements. Several of the walls showed some signs of distress (i.e. settlement
and longitudinal tension cracks along the retained fill, slight out-of-plane bulging). It is
noteworthy to mention that some of those structures were shaken with acceleration levels
almost twice the values they were designed against (Sandri 1997).
539
There were 23 Reinforced Earth structures built with steel strips and concrete facing panels
(Frankenberger et al. 1996; Sitar et al. 1997). Tallest of these structures was 17 meters. The
walls performed well. Some panels separated causing spillage of material and some minor
cracking was observed. Only one 16 meter wall experienced some problems, wall face
bulging 46 centimeters, 3% of the wall height (Sitar et al. 1997).
HYOGOKEN-NANBU EARTHQUAKE (KOBE), JAPAN, 1995
This was a magnitude 6.9 earthquake that caused significant damage to Kobe and
surroundings. Several case studies of reinforced soil structure performance are reported by
researchers from Japan (Kobayashi et al. 1996; Nishimura et al. 1996; Tatsuoka et al. 1995,
1996a; Tatsuoka et al. 1996b; Tatsuoka et al. 1998).
Performances of four geogrid reinforced walls with heights between 3 to 8 meters are
reported. These walls support the railway embankment and also serve as bridge abutment at
several locations. These structures were composed of a monolithic cast-in-place facing and
relatively short reinforcements. (Tatsuoka et al. 1995, 1996a; Tatsuoka et al. 1996b; Tatsuoka
et al. 1998). These walls were designed with the pseudo-static method using a seismic
horizontal coefficient of 0.2. Peak ground accelerations at these sites reached up to 0.8g.
Several nearby conventional retaining walls and houses suffered heavy damage. Three of the
walls performed very well and exhibited no visual damage. The cross section of one of the
walls along Japan Railway Tokaido Line is shown in Figure A-3. There is an electric supply
frame located behind the wall and the 2.2 meter wide circular foundation of this frame
embedded deep behind the retaining structure. A detailed description of this geogrid-
reinforced retaining structure including construction details, monitoring data during and after
construction is given in Kanazawa et al. (1992). Additionally, full-scale experimental studies
performed to investigate the effect of the embedded foundation are reported in Tamura et al.
(1992). A plan view and cross section of other geogrid-reinforced retaining walls along the
same railroad line are given in Figure A-4. These are utilized as bridge abutments. Geogrid-
reinforced soil structures at this site at Amagasaki, Japan and at two other sites performed
very well. However, the fourth wall (Tanata wall) experienced some damage. The wall
540
moved about 30 cm horizontally as seen in Figure A-5. The wall also suffered facing panels
spalling and cracking. This wall was built with relatively short reinforcements.
In addition to the above geogrid-reinforced walls, ten other structures were identified
(Nishimura et al. 1996). Heights of these walls ranged between 4 to 11 meters. These
structured performed very well with no signs of damage except two of the walls where some
minor cracks and facing separation were observed.
Mechanically stabilized earth walls built with steel strips (i.e. Reinforced Earth) also
performed well. Kobayashi et al. (1996) reports 124 Reinforced Earth structures within 40
km of the earthquake source. Twenty-one of these were within the area of severe shaking.
Most of these structures performed very well. However, some signs of damage were
observed at three of these walls. One of walls that was built as part of the approach
embankment is shown in Figure A-6.
The wall was 9 meters high at its maximum and was designed with a seismic coefficient of
0.15. Earthquake caused the top of the wall move outward about 15 cm. Similar damage
patterns were observed at 2 other Reinforced Earth walls. Tatsuoka et al. (1996a) indicates 66
of these structures were identified within 70 km of the earthquake source. Obviously
Reinforced Earth wall inventory from both reports overlap (Kobayashi et al. 1996; Tatsuoka
et al. 1996a) and it is not possible to identify cases that are not reported in detail. Tatsuoka et
al. (1996a) describes the same three walls that suffered damage and asserts the similar
damage assessments.
Seven slopes/excavations stabilized by soil nailing were reported (Tatsuoka et al. 1995). No
noticeable deformation or signs of damage were observed at these structures except at one
excavation supported by soil nailing. Soil nailing was used to support a 14-meter deep
excavation. The excavation was at a late stage at the time up to the earthquake and was
backfilled leaving a free face of about 4 meters. Comparison of before and after earthquake
inclinometer measurements indicate that top of the wall deformed about 3 millimeters.
541
CHI-CHI EARTHQUAKE, TAIWAN, 1999
This was a magnitude 7.6 event and caused significant damage to structures in urban areas.
Several studies have been performed on the performance of reinforced soil structures during
this earthquake (Chen et al. 2000; Huang 2000; Ling et al. 2001). Geosynthetic reinforced
walls with modular facing blocks are commonly used in Taiwan and this earthquake
provided the first opportunity to test the performance of this type of MSE structures. Several
failures have been reported where mechanically stabilized embankments and reinforced soil
slopes experienced significant damage and total collapse.
Four sites with modular-block geosynthetic-reinforced retaining walls were identified.
(Huang 2000; Ling et al. 2001). At one of these sites modular-lock geosynthetic-reinforced
retaining walls were utilized to retain excavated slopes for a housing development project.
Large cracks and settlements were observed along the slopes indicating global stability
problems with the slopes in the area. Several reinforced concrete retaining walls and
unreinforced modular-block walls at the site collapsed and/or got damaged. The height of the
reinforced soil structures at the site varied from location to location, reaching 5 meters at its
highest. Parts of these modular-block geosynthetic-reinforced soil structures experienced
damage. At one location blocks near the top of the walls displaced outward and at another
section the blocks fell apart. It was observed that the walls at this site were built using
geogrid reinforcements and good quality backfill.
At another site along Ta Kung Roadway 129 a section of a geosynthetic-reinforced wall with
modular block facing collapsed and other sections suffered heavy damage. This wall was
built with geogrid reinforcements and silty sand backfill. The wall was 3.4 meters at the
collapsed section. A cross section of the collapse is shown in Figure A-7. In addition to the
section that collapsed, other sections of the wall also suffered significant damage where
blocks moved out-of-alignment and separated causing backfill material to spill out. The
largest bulging displacement was observed at about 1.6 meters from the bottom of the wall
(about 1/2 to 1/3 of the wall height). Geogrid reinforcements were observed to be torn at
location of connection pins. The cross section of the wall at this part of the wall and a sketch
of the observed deformations are shown in Figure A-8. A 7.5-meter high conventional
542
reinforced concrete retaining wall supports the other side of the highway embankment and
suffered only minor damage. This RC retaining wall was designed by the same organization
that built the reinforced soil structure. Both structures were built around the same period of
time (Huang 2000). Peak ground accelerations during the earthquake were in excess of 0.4g
in the area whereas seismic provisions require a seismic coefficient kh=0.115 for this area.
Two short walls located near a stadium suffered some damage. One of the walls was 2 meters
high and facing blocks separated. It is noted that this was due to the movement of the
foundation of lamp-posts located behind the wall. The other wall was 3 meters high and
collapsed. This failure was attributed to short reinforcements. Longitudinal cracks were
observed behind the wall.
At the fourth site a reinforced soil structure suffered some damage. This is a geosynthetic-
reinforced retaining wall composed of three stacked walls. The height of the wall is not
mentioned in the paper, however from the given photograph it appears the wall is composed
of 33 modular blocks making it about 6.5 meters high (given each block is 20 cm). Part of the
wall that was not reinforced collapsed.
A 40-meter high reinforced slope collapsed during the earthquake. Geogrids were used as
reinforcements and wrapped around as facing. On-site silty clay was used as backfill material
(Huang 2000; Ling et al. 2001). The total height of the slope is 80 meters, 40 meters of that
composed of 4 reinforced sections each 10 meters high as shown in Figure A-9. This slope
failed in 1994 immediately following completion and large deformations were observed in
1996 following repair. In addition to this collapse-repair sequence, parts of the unreinforced
slope beneath the reinforced slope were strengthened with reinforced concrete frame and tie-
back anchors. All these complexities make it a difficult case to assess the mechanism that
lead to collapse during the earthquake.
This slope was designed using a pseudo-static seismic coefficient kh=0.15 and a pore
pressure coefficient ru=0. Slope stability analyses using these values in the design phase
yielded a Factor of Safety, FS=1.1. Stability analyses performed after the earthquake, using
the same parameters resulted in FS=1.0 (Huang 2000). In any case this slope exhibited
stability problems immediately following construction and conventional stability analyses
543
yield a marginal factor of safety. Studies have been performed on this case study (Chen et al.
2000; Holtz et al. 2001). However, limitations on the pre-earthquake conditions prevent these
studies to be conclusive about the collapse.
In another case, a 35 meter high reinforced slope performed well. Geogrids were used as
reinforcements and wrapped around as facing (Ling et al. 2001). Detailed description of this
reinforced slope are given in Chou et al. (1994).
NISQUALLY EARTHQUAKE, WASHINGTON, 2000
A mechanically-stabilized earth wall supporting a hotel parking lot collapsed following the
earthquake (Kramer et al. 2001). This structure was built with geogrid reinforcements and
modular-block concrete. Parts of this wall experienced problems following construction and
were repaired. It is not clear though if the failed section was actually part of the repaired
segments or not. It is also possible that soft foundation soils contributed to this collapse.
Extended Stay Hotel MSE Wall- Seattle: (Wall Failure; Geosynthetic-Reinforced MSE Wall; Weak Foundation)
The Extended Stay Wall in Tumwater, WA is a geosynthetic reinforced MSE structure with
concrete block facing (Kramer et al. 2001). The site is approximately 27 km from the
epicenter of the 2001 Nisqually Earthquake and the peak ground accelerations in the area
reached about 0.15g. The wall is 4.5 m high with 3.5 m-long geogrid reinforcement. The wall
supports a backfill slope with a 2:1 grade. Sections of the wall collapsed during the
earthquake. It is highly probable that the poor subsoil conditions were the major factors that
caused the collapse.
Construction reports indicate that problems with the foundation conditions were noticed prior
to and during construction (Geo-Group Northwest, 2000). Some segments of the wall were
demolished, soft foundation soils removed and rebuilt. It is not clear which segments
suffered heavy damage during the earthquake and if the repair efforts were effective in
earthquake performance.
544
Costco MSE Wall – Seattle: (Large Movements; Geosynthetic-Reinforced MSE Wall; Inadequate Base Width)
The Costco MSE Wall near Tacoma, WA is a geosynthetic-reinforced MSE structure with
concrete block facing that retains the parking lot area of the Costco Department Store. The
site is approximately 30 km from the epicenter and the peak ground accelerations in the area
reached about 0.07g during the event.
The wall is about 5.5 meters high at the maximum section and the reinforcement length at
this section is 2.8 meters, giving a width-to-height ratio of 0.50 (Geo-Engineers 1999). At
this section, the wall retains a level backfill. A sloping backfill (H:V 2:1) is retained at some
sections. Design drawings indicate that the width-to-height ratio at these locations is typically
0.70. These width-to-height ratios lower than those generally recommended (0.70). Geogrid
bars were used as reinforcing elements and were placed a 1.0 to 1.5-m vertical spacings.
The wall remained intact after the earthquake. Lateral displacements due to shaking resulted
in a wedge-type sliding behind the wall. This movement produced large cracks at the crest of
the sloping backfill and retained soil with horizontal offsets of up to 25 cm and vertical
offsets in the range of 2 to 3 cm. The exact displacement pattern of the structure that led to
this pattern is not known. Typically in such cases the wall may move laterally as a rigid block
or it may deform undergoing internal straining, tilting etc. Some out-of-plane bulging was
observed at lower elevations of the wall at some sections.
This is an important case because it demonstrates how these structures can undergo large
displacements before actually collapsing. This is consistent with our reconnaissance
observations and recent dynamic analysis results of the Arifiye RE wall associated with the
1999 Kocaeli (Turkey) Earthquake. This wall should also provide some insight into the
effects of sloping backfill, geosynthetic structural elements (as opposed to more rigid steel
strips, etc.), and width-to-height ratio.
545
Figure A-1. Cross section of La Honda reinforced slope (Collin et al. 1992)
Figure A-2. Inclinometer data showing before and after earthquake measurements (Collin et al. 1992)
Deflection (cm)
0 1 2 3 4 5 6
Dep
th (m
)
0
5
10
15
20
Before EQSept. 1989
After EQOct. 1989
Tensar geogrid reinforcement
12" permeable material lined in filter fabric
rock slope protection
Horizontal drainage outlet
5 % 4.6 m
9.1 m
3 meters minimum
546
Figure A-3. Cross-section of a geogrid-reinforced wall along the railway (Tatsuoka et al. 1996a)
547
Figure A-4. Plan and cross-section of geogrid-reinforced retaining wall bridge abutments (Kanazawa et al. 1992)
548
Figure A-5. Cross section of geogrid-reinforced wall that suffered some damage (Tatsuoka et al. 1996a)
Figure A-6. Cross section of the Reinforced Earth wall that suffered some damage (Tatsuoka et al. 1996a)
549
Figure A-7. Cross section of the collapsed geosynthetic reinforced modular block wall (Huang 2000)
Figure A-8. Large deformations at the geosynthetic reinforced modular block wall (Huang 2000)
550
Figure A-9. Cross section of the failed reinforced slope (adapted from Huang 2000)
After Failure
Original Slope
Reinforced Slopeas Designed
Sv = 1 m
15m
10
m
10m
26
m
10m
10
m
Tult = 104 kN/m L = 13 m
Tult = 60 kN/m L = 10 m
Tult = 60 kN/m L = 7 m
Tult = 60 kN/m L = 4 m
/ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 600 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages false /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages false /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown
/Description >>> setdistillerparams> setpagedevice